Skip to main content

SilaiSathi × Graphify

Offline AI workstation for Indian tailors with a Graphify code radar to trace dependencies and calculate pre-change blast radius.

About the project

SilaiSathi is an offline-first AI workstation designed for independent Indian tailoring shops and boutiques. It helps turn Hindi, Hinglish, and English customer requests into structured tailoring job information, with human review, local persistence, deterministic financial calculations, and printable work slips.

For Hacktoberfest Hack Day, SilaiSathi integrates Graphify into its developer workflow. The generated code knowledge graph helps developers explore module relationships, inspect symbols, trace dependencies, and estimate the potential impact of code changes before editing critical functionality.

The integration includes a developer-facing architectural radar and graph inspection tools. Its purpose is to make code navigation more informative than flat text search when investigating relationships across modules.

The application prioritizes local processing and privacy. Its AI runtime uses the locally configured Ollama model, while Graphify is used for codebase analysis.

Source code: https://github.com/dikshant363/silaisathi

Built with

pythonstreamlitsqlitepydanticollamapytestruff

AI tools

Coded withAntigravity

Prize Categories Entered

  • Deepmind Google

    Best Use of Gemma 4

    Submitted

    SilaiSathi integrates Google Gemma 4 (gemma-4-26b-a4b-it via the Google GenAI SDK) as an optional cloud inference provider alongside offline Ollama. Gemma 4 parses complex, multilingual Hinglish tailor requests and conversational voice logs into structured JSON tailoring job cards with missing measurement detection, while all financial and ledger calculations remain strictly quarantined in deterministic Python.

  • DigitalOcean

    Best Open-Source AI Project

    Submitted

    SilaiSathi combines a locally run open-weight language model (Qwen 3.5 4B via Ollama) with an offline-first workflow for independent Indian tailoring shops. Its Graphify integration adds graph-based code understanding to the open-source developer workflow, providing an offline AST architectural radar under an MIT open-source license.

  • Event prize

    Best Use of Graphify

    Submitted

    SilaiSathi uses Graphify to generate an offline knowledge graph of its codebase and power an architectural radar for exploring symbols, dependencies, and potential change impact. The integration demonstrates how graph-based code understanding helps developers investigate cross-module relationships and debug issues systematically before code mutations.